Product Manager, Compute Operations

Fluidstack•New York, NY

4d•$245,000 - $288,000

About The Position

Fluidstack is building civilization-scale infrastructure for AI, focusing on delivering gigawatt-scale compute faster than anyone else. This involves rethinking every layer of the stack, from acquiring power and designing data centers to operating them. The company is looking for individuals who care deeply about this problem space and are driven by extreme ownership, velocity, first principles, and a passion for the frontier of AI. The Compute Production team specifically builds the data foundations and internal products necessary to bring AI compute online and maintain its operation. They tackle problems such as building an ontology for compute production, consolidating fragmented operational data into a single source of truth, shipping product surfaces to enhance the efficiency of compute production teams, and applying LLM-era tooling to operational workflows. This role will own the product direction for the team's ontology and data products serving compute production. This includes defining the roadmap, setting priorities, and establishing success metrics. The Product Manager will be responsible for transforming complex compute production domains into clear data models, specifications, and product surfaces that engineers can utilize. They will drive technical tradeoffs with engineers, manage the end-to-end product lifecycle from problem definition to delivery, and conduct direct discovery with stakeholders to understand their needs and validate product impact. Defining and tracking metrics to demonstrate product leverage and cutting or redirecting ineffective work are also key aspects of the role.

Requirements

3+ years of technical product management, ideally on data, platform, or ontology products
A track record of shipping products with measurable impact, and the numbers to show it
You can reason about system design and hold your own with engineers on technical tradeoffs, not just relay their conclusions
Strong written communication: crisp specs, and prioritization rationale sharp enough that stakeholders who disagree still understand why you chose what you chose
You're comfortable in ambiguity and know how to impose structure where none exists, turning an unmapped operational domain into a model others can build on

Nice To Haves

exposure to compute or datacenter operations, infrastructure systems, or building LLM-powered applications

Responsibilities

Own product direction for the team's ontology and data products serving compute production: the roadmap, the priorities, and the success metrics that justify them
Turn messy compute production domains into data models, specs, and product surfaces clear enough that engineers can ship against them without a second round of discovery
Drive technical tradeoffs with engineers and ship end to end, from problem definition through delivery
Run direct discovery with compute production stakeholders to understand their domain, then validate that what you built actually changes how they work
Define and track the metrics that prove the product is delivering real leverage, and cut or redirect work that isn't